Document Type

Technical Report

Publication Date



Categorical and regression tree analysis; dissolved phosphorus, Red River Basin, water monitoring


The Red River is a trans‐boundary, multi‐jurisdictional basin, where water‐quality standards often change at the state lines. The state agencies with USEPA Region VI focused resources to organize water‐quality data from within this basin and have it statistically analyzed to evaluate the relationships between nutrients and sestonic chlorophyll‐a (chl‐a). There were 152 sites within the Red River Basin that had nutrient and sestonic chl‐a data, and these sites were narrowed down to 132 when a minimum number of observations was required. Sestonic chl‐a levels increased with increasing nutrient concentrations; these significant regressions were used to predict nutrient concentrations at 10 µg chl‐a L⁻¹. Total nitrogen (TN) and phosphorus (TP) concentrations (at 10 µg chl‐a L⁻¹ ) varied across the Red River Basin and its eco‐regions from 0.10‐0.22 mg TP L⁻¹ and 0.75‐2.11 mg TN L⁻¹ . Nutrient thresholds were also observed with sestonic chl‐a at 0.14 mg TP L⁻¹ and 0.74 mg TN L⁻¹ using categorical and regression tree analysis (CART). CART analysis also revealed that hierarchical structure was important when attempting to predict sestonic chl‐a from TN, TP and conductivity. The ranges of TN and TP concentrations that resulted in chl‐a concentrations which exceeded 10 µg chl‐a L⁻¹ were similar in magnitude to the threshold in TN and TP that resulted in increased sestonic chl‐a. This corroborating evidence provides useful guidance to the states with jurisdiction within the Red River Basin for establishing nutrient criteria, which might be similar when the Red River and its tributaries cross political boundaries.

Report Number